Question

我正在尝试找到一种方法，可以通过多个类别对人们的分数进行排名。该表包含每个玩家的列表以及他们参加的事件和得分。

查询需要根据每个事件中的得分为每个玩家分配等级。（在事件1和事件2中分配的等级应该完全分开。）

如果某个事件中缺少玩家，则应为他们分配（该类别中的玩家总数+ 1）的排名

我可以对1个类别进行基本排名。很简单但是我什至不知道该怎么做。

表中的数据

+--------+-------+---------+
| Name   | Score | Event # |
+--------+-------+---------+
| Kevin  | 73    | Event 1 |
| George | 69    | Event 1 |
| Henry  | 70    | Event 1 |
|        |       |         |
| George | 45    | Event 2 |
| Kevin  | 65    | Event 2 |
| Henry  | 65    | Event 2 |
| Daniel | 50    | Event 2 |
+--------+-------+---------+

预期回报

+---------+--------+------+
| Event # | Name   | Rank |
+---------+--------+------+
| Event 1 | Kevin  | 1    |
| Event 1 | Henry  | 2    |
| Event 1 | George | 3    |
| Event 1 | Daniel | 4    |
| Event 2 | Kevin  | 1    |
| Event 2 | Henry  | 1    |
| Event 2 | Daniel | 3    |
| Event 2 | George | 4    |
+---------+--------+------+

请注意，排名并不密集！我使用的是MySql 5.7。

实际上，我有50多个活动和数百人。

这是我排名1项赛事的条件

SELECT
        IF( SCORE =@_last_rank,@curRank:=@curRank,@curRank:=@_sequence) AS RANK,
        NAME,
        EVENT,
        @_sequence:=@_sequence+1,@_last_rank:= SCORE
        FROM (SELECT * FROM database WHERE EVENT = $event) p, 
        (SELECT @curRank := 1, @_sequence:=1, @_last_rank:=0) r
        ORDER BY  RANK, NAME

Answer 1

此sql语句

Select eventno,Name,rank 
From 
 (Select eventno,Name,
            CASE 
              WHEN @event = eventno THEN @rownum := @rownum + 1
              ELSE 
                  @rownum := 1
             END as rank,
             @event := eventno
      From table1 t
      JOIN (SELECT @rownum := 0) r
      JOIN (SELECT @event := (Select Eventno from table1 Limit 1)) s
  ORDER BY t.eventno,t.score) x;


  CREATE TABLE table1
  (`Name` varchar(6), `Score` int, `Eventno` varchar(7)) 

  INSERT INTO table1
(`Name`, `Score`, `Eventno`)
VALUES
('Kevin', 73, 'Event 1'),
('George', 69, 'Event 1'),
('Henry', 70, 'Event 1'),
('George', 45, 'Event 2'),
('Kevin', 65, 'Event 2'),
('Henry', 65, 'Event 2'),
('Daniel', 50, 'Event 2')

给出以下结果：

  eventno   Name    rank
  Event 1   George  1
  Event 1   Henry   2
  Event 1   Kevin   3
  Event 2   George  1
  Event 2   Daniel  2
  Event 2   Kevin   3
  Event 2   Henry   4
  7 rows

Answer 2

以下方法避免了（不可靠的）使用用户变量和高度复杂的查询。它使用临时表并利用MyISAM的每组自动输入功能，以每组生成行号。为了从行号中获得排名，使用了带有MIN(row_number)的GROUP BY查询。

drop temporary table if exists tmp_scores;
create temporary table tmp_scores(
  rn int auto_increment,
  event varchar(50),
  name varchar(50),
  score int,
  primary key (event, rn)
)engine=myisam;

insert into tmp_scores(event, name, score)
  select e.event, n.name, t.score
  from (select distinct t.event from mytable t) e
  cross join (select distinct t.name from mytable t) n
  left join mytable t on t.event = e.event and t.name = n.name
  order by e.event, t.score is null, t.score desc;

drop temporary table if exists tmp_rank_by_score;
create temporary table tmp_rank_by_score(
  event varchar(50),
  score int,
  rank int,
  primary key (event, rank)
)engine=myisam;

insert into tmp_rank_by_score(event, score, rank)
  select event, score, min(rn) as rank
  from tmp_scores
  group by event, score;

select s.event, s.name, s.score, r.rank
from tmp_scores s
join tmp_rank_by_score r on r.event = s.event and r.score <=> s.score
order by r.event, r.rank;

drop temporary table if exists tmp_rank_by_score;
drop temporary table if exists tmp_scores;

结果：

| event   | name   | score | rank |
| ------- | ------ | ----- | ---- |
| Event 1 | Kevin  | 73    | 1    |
| Event 1 | Henry  | 70    | 2    |
| Event 1 | George | 69    | 3    |
| Event 1 | Daniel |       | 4    |
| Event 2 | Kevin  | 65    | 1    |
| Event 2 | Henry  | 65    | 1    |
| Event 2 | Daniel | 50    | 3    |
| Event 2 | George | 45    | 4    |

db-fiddle

通过多个类别对表格进行排名

2 个答案: